Damage to the nervous system often results in altered movement control, leading to loss of function. The NIH estimates that movement-related neurological disorders affect millions of Americans each year. For many neurological disorders, rehabilitation therapy is a main treatment. In order to optimize rehabilitation techniques, a quantitative approach that pinpoints specific deficiencies and targets them with long-term intervention is needed. Thus, we will apply biomechanical models, robotic devices and novel control strategies for optimizing both compensation and learning approaches for improved movement control. This project focuses on poor motor coordination, which is a ubiquitous finding in people with damage to the nervous system. Incoordination leads to poor trajectory and targeting control, and is most distinctly related to cerebellar dysfunction. The fundamental mechanism of cerebellar incoordination will be investigated by comparing human performance to computational models, and then by developing robotic control strategies that either compensate for motor impairments or optimize practice-dependent learning. Both approaches are needed since short-term learning mechanisms can be inefficient or absent in individuals with cerebellar damage, making compensation the best option for some people. A robotic exoskeleton device, the KinArm, will be used to acquire behavioral data during reaching tasks performed by control and cerebellar subjects, and dynamic models of the human arm will be used to determine the source of the differences between control and cerebellar data. Specifically, the dynamic models of subjects will be used to simulate the effects of misestimation of limb dynamics and timing delays. The parameters that best explain behavior will be used to inform a rational control strategy for robot-assisted rehabilitation for ataxic populations. Adaptation and compensation methods will be designed that provide assistive and/or resistive forces to help subjects achieve normal movement patterns. A pilot study in which cerebellar patients use these methods will provide design guidelines for future rehabilitation robotics development. The long-term goal of this work is to design and produce take-home devices that are customized to either compensate for an individual's deficit or to facilitate the learning of a new motor pattern. We plan to extend this methodology to a broad range of patient populations. This project lays the foundation for novel home therapies by identifying strategies a robot could use to normalize movement control of people with cerebellar damage. PUBLIC HEALTH RELEVANCE. Movement disorders commonly occur following neurological damage, affecting the activities of daily living for millions of Americans each year. Therapies and assistance methods using rehabilitation robots are promising techniques for improving the short- and long-term health care of these patients, by lowering the cost of treatment, enabling more effective methods for practice-based rehabilitation, and providing "smart" orthoses for recovery of normal movement function in populations for whom adaptation is not possible.